Decomposition-Based Multi-Step Forecasting Model for the Environmental Variables of Rabbit Houses

Area of Science:

• Agricultural Engineering
• Environmental Monitoring
• Time Series Analysis

Background:

• Traditional methods for forecasting rabbit house environmental parameters are inadequate due to ignoring sequence coupling.
• Accurate prediction of environmental variables like temperature, humidity, and CO2 is crucial for animal welfare and management.

Purpose of the Study:

• To develop a decomposition-based multi-step forecasting model for rabbit house environmental variables.
• To enhance prediction accuracy and provide timely control decision-making for rabbit housing environments.

Main Methods:

• The proposed model decomposes non-stationary time series into trend, seasonal, and residual components using the STL algorithm.
• LSTM and Informer models are employed to predict the trend and residual components, respectively.
• The final prediction is obtained by combining the predicted components with the seasonal component, utilizing a multi-input, single-output approach.

Main Results:

• The model significantly improved prediction accuracy for temperature and humidity in rabbit houses.
• The Informer model's performance was sensitive to encoder and decoder input sequence lengths.
• While effective for temperature and humidity, the model showed less improvement for carbon dioxide concentration prediction.

Conclusions:

• The decomposition-based multi-step forecasting model effectively extracts coupling relationships among correlated time series.
• The model demonstrates strong capabilities for multivariate multi-step prediction of non-stationary time series in rabbit housing environments.
• Further refinement may be needed to optimize carbon dioxide concentration forecasting.